Flush Air Data System Fault Detection Using Pressure Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fault detection and isolation (FDI) methods in Flush Air Data Systems (FADS) are cumbersome and rely on inverse models, requiring extensive validation of statistical characteristics, which complicates onboard implementation and can lead to loss of control and mission failure due to faulty pressure measurements.
Innovation Solution
A novel FDI methodology that directly works with input pressures, classifies pressure ports as outer, inner, and centre, and uses nearest neighbour checks with predefined thresholds to detect and isolate faulty ports, providing redundancy to tolerate blockages and sensor failures, ensuring reliable air data output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inverse models with neural networks are used for fault detection, then measurement precision can be improved, but device complexity increases significantly
Solution Approach 1:
The fault detection system is segmented into multiple independent linear models, each responsible for a specific pressure port. Instead of using a single complex inverse model for all ports, the system divides the problem into smaller segments that can be processed independently and efficiently
Solution Approach 2:
The patent replaces complex computational mechanics (neural networks and inverse models) with simpler linear models. This substitution maintains adequate fault detection capability while dramatically reducing computational complexity and making the system suitable for onboard implementation
2Reliability
If extensive validation of statistical characteristics is performed, then reliability of fault detection improves, but loss of time increases
Solution Approach 1:
Statistical validation of the linear models is performed beforehand during system development and testing, rather than during real-time operation. This preliminary action ensures model reliability is established in advance, allowing rapid fault detection during actual flight operations without time-consuming validation
Solution Approach 2:
The system changes from requiring extensive statistical validation during operation to using pre-validated linear models with simplified real-time checks. By transforming the validation process into a preliminary parameter setup phase, the system achieves both reliability and real-time performance
3Reliability
If redundant pressure ports are added to tolerate failures, then system reliability improves, but device complexity increases
Solution Approach 1:
The patent merges the functionality of fault detection and redundancy management into a unified linear model framework. Multiple pressure ports including redundant ones are integrated into the same simple computational structure, allowing the system to tolerate failures without requiring separate complex fault management subsystems
Data Source
Figure 1a~2
Figure 3~4a
Figure 4b~4c
AI summary
This invention relates to a methodology for detection and isolating in real time blocked pressure ports (outer & inner ports) and/or faulty pressure measurements in Flush Air Data Systems (FADS). Pressure ports lie along the periphery of the outer and inner circle while the inner most pressure port is located at the centre of the nose cap. The max. pressure difference that can occur between two adjacent ports for different flight conditions of Mach number, angle of attack and side slip angle is established initially through extensive simulations. During flight, each pressure port reading is compared against predefined thresholds with its nearest two neighbours in the corresponding circle. The faulty port pair is further isolated by detecting significant changes than normally expected over a time period of one minor cycle computation time of the processor. Two sets of thresholds are used, for Mach > 1 and Mach<1.